Playing the Game, Inscribing the Landscape

The archaeology of mancala games in southern Malawi

Menno Welling Research School CNWS, Leiden University

1 “In play there is something ‘at play’ ” Johan Huizinga in Homo Ludens (1938)

2 Mancala Games

• Games with undifferentiated counters being spread over a board with ‘cups’. • Aim: to capture the counters of your opponent by spreading yours one by one • Stone/wooden/clay boards with 2, 3 or 4 rows • Wide variety of rules E.g. captured counters are either taken out or appropriated and resown • Found in Africa, all of Southern Asia and Brazil/Caribbean

3 Distribution

From Popova 1977

In terms of usage three-row games can be considered compressed two-row games. The two-row game is wide-spread in southern Asia, the northern half of Africa and the African Diaspora in the Caribbean. Four-row mancala has been thought to be restricted to southern, more or less Bantu-speaking Africa (Popova 1977). The association of Bantu speakers with four-row mancala is however not entirely correct (See Towshend 1979). Two-row mancala has for instance been reported in Angola, four-row mancala in southern Sudan, but most significantly, rock cut mancala games consisting of four rows abound in the old city of Petra.

4 Origin

3500 year old Stone board, Memphis, Egypt from Petrie (1927)

Some of the oldest mancala games go back to the New Kingdom in Egypt. Such as the excavated by Petrie and those carved in the roof of the temple of Set I which have two rows of six cups.

5 Anthropological studies

• Origin and diffusion (Townshend 1979) – E.g. Zanzibar: • Bao KiSwahili (4x8, captured counters reinserted) • Bao KiArabu (4x7/8, captures taken out) • Bao KiSwahili original, • Bao KiArabu from interior (Lake Malawi area) • Mathematical aspects • Cultural aspects: – Rainmaking – Divination – Politics – Judiciary system

There have been many anthropological, or quasi anthropological studies of mancala. Most present typologies and diffusions. Townshend speculation that bao kiArabu could originate from Lake Malawi is however unfounded as no 4x8 boards are known to predate the bao kiSwahili in the area.

6 Archaeology?

In the archaeological literature, rock cut mancala games are often just mentioned but sometimes not even that –particular in areas where they are ubiquitous. The game boards are rarely adequately described late alone analyzed. Of course there is the problems of dating of rock surfaces. Moreover, the social context is often lacking, let alone that a reconstruction of the rules is a possibility. Yet these games can be fruitfully subjected to archaeological analysis. Even with regard to rock cut games that can be found troughout east, central and southern Africa from Egypt down to Angola , Namibia and South Africa (also see Walker 1991). Particularly, large concentrations can be found in the northern South Africa/southerns Zimbabwe area. The ones at Schroda would probably be the oldest.

7 Uganda from Wayland (1936)

Namibia from Schwertz (1975)

8 Mapungubwa, SA from Huffman (2005)

Aksum

Thimlich, Kenya photo V. Emmerson

9 A Case Study from Southern Malawi

In the following discussion of rock cut mancala boards in Southern Malawi, I will make use of the concept of belonging as is currently in vogue in ethnographies of migrant communities in the globalizing world. I will particularly follow it usage by Fortier in her study of an Italian community in central London.

10 Belonging “Imagining a community is both that which is created as a common history, experience or culture of a group –a group’s belonging - and about how the imagined community is attached to places – the location of culture” (Fortier 1999)

Situated belonging as the outcome of ‘iterated performance’.

11 map

12 Southern Malawi

• Nyanja/Mang’anja population: the descendents of the Maravi states, 1200- 1700AD. • Nyanja/Mang’anja are matrilineal and uxorilocal. • Known for the secret male society called nyau performing at initiations and funerals. • Yao, Ngoni incursions 19 th Century. • Lomwe, Tonga immigration from 1890.

13 Nyau

14 Mancala in Malawi

• In the past various games: Msuwa , Fuwa , Mjombwa , Nsolo , Msuwo , Chiana etc. • All: 4 row, but no 4x8. In all captures taken out, except in one children’s version. • Replaced by Bao kiSwahili from Zanzibar . • Brought by Yao groups, late 19 th century (Sanderson 1913)

15 Fieldwork 2005

16 TO43 R. Henderson Residence

17 TO44 D. Henderson Residence

18 TO45 Tobola

19 TO46 January Village

20 TO47 January School

21 22 Overview

• 5 sites • 10 games of 4x8, some with repair • Probably sequential, not synchronous usage of games at any one site • Mawudzu/longwe ware at 3 sites (poor visibility at 2 sites) • Pottery located only at one side of the rock outcrops • All sites close to water source

23 Conclusions

• 4x8 games dating to c.1200-1700: pre-Yao immigration: • While Bao KiSwahili is introduced from Zanzibar by Yao groups in the late 19 th century, Bao KiArabu can actually have traveled in opposite direction.

24 Conclusions cont.

• The games located at the village edge. • Against Werner 1906:“The four rows of holes can be made on any bit of smooth ground, and one often sees them in the bwalo of a nyanja village, where the men sit smoking and gossiping and weaving baskets ” • Considering land tenure: stone boards likely to be collectively owned. • Loci of male performance of group identity, thus in incorporating the incoming affines.

The games were located at the edge of the village. Situated between bush (nature) and village (culture) the rocks make an ideal location for play. Werner’s statement, if reliable, that they can be found in the bwalo , or village square could indicate a regional or historical difference. Yet, games scooped out of the ground could have coexisted with the rock cut games. The materiality of the game boards is however of prime importance. Stone boards are more durable than wooden board. Moreover they are immovable. Not only does this make them more likely to have collective rather than private ownership, it makes the rock outcrops suitable loci of group identity. Not surprisingly throughout Malawi rock formations can be found that function as loci of memory and belonging (e.g. Kaphirintiwa, Mbona). The sequential usage of the games on multi-game rock surfaces point to this group performance. Being there was more important than actually playing. Paraphrasing Fortier I would argue the mancala games are “[p]ractises of group identity [that] are about manufacturing cultural and historical belongings which mark out terrains of communality that delineate the politics and social dynamics of ‘fitting in’.” In that sense it is a secular analogue to the male society of nyau. In the process of playing the came. The gender component to Malawi mancala is further highlighted in the fact that the two common versions of Bao KiSwahili which in Swahili are called bao yokhoma and bao yawana , in Malawi are often referred to as achimuna and achikazi : the male (difficult) and female (easy) version. In the end the rock cut mancala games become loci of nested male identity. In the process the incoming men inscribe the landscape that is not their own but that of their wives.

25 Acknowledgement

• Field assistant B. Kayira of the Department of Antiquities. • Laiton Kasonjora for showing some of the sites. • Dottie and Ralph Henderson for permission to access their land. • Ted and Bette Best for making it possible to make this presentation here.

26